Dust vs GPT Researcher (2026): which should you pick?

Pick Dust if

• Fits if your company wants agents that answer from its own docs and chats
• Works for organizations that need their data kept in the US or the EU

Look elsewhere if

• Tight if you're staying free, as the Free seat is 500 credits once and never refills
• Not for you if you need more than 3 data sources, 5 MCP servers or 5 spaces without an Enterprise deal
• Skip it if you need audit logs, SCIM or custom data retention, which sit on the custom-priced Enterprise plan
• Pricey if heavy users need Max, at $120 a seat a month billed yearly or $150 monthly
• Watch out: credits don't roll over, prices exclude VAT, and there's no phone or desktop app

Pick GPT Researcher if

• Fits if you can run Python or Docker and want cited research reports
• Works for developers who want to choose their own search sources and models

Look elsewhere if

• Not for you if you want a hosted service, since there is none
• Skip it if you do not want a command line, as setup needs Python 3.11 or newer or Docker
• Watch out: you pay your LLM and search providers directly
• Too light on support: managed deployments and SLAs are by contacting the maintainers, with no price

Dust

• The Free seat is a one-time allowance of 500 credits, not a monthly quota; once it is used, free users are prompted to request an upgrade.
• The self-serve Business plan covers teams of up to 100 and is limited to 3 data source connectors, 5 remote MCP servers and 5 spaces.
• SCIM provisioning, audit logs, custom data retention and single-tenant deployment require Enterprise.
• Monthly credits reset each billing period without rolling over, and listed prices exclude VAT.

GPT Researcher

• There is no hosted service; you run it yourself with Python 3.11 or newer, or Docker.
• You pay your LLM and search providers directly; the site gives a few cents to a few dollars per deep run as a ballpark.
• Managed deployments, private-corpus retrievers and SLAs are handled by contacting the maintainers, with no listed price.

Dust

Agents with company context

Agents draw on connected company data from sources such as Slack, Notion, Google Drive, GitHub, Salesforce and Zendesk to answer questions and take action.

Model choice per agent

All plans include 20+ models from OpenAI, Anthropic, Google, Mistral and DeepSeek. The model is chosen per agent, and none is locked to a higher plan.

Schedules and triggers

Multi-agent workflows can run on a schedule or in response to events. Dust also lists Zapier, Make, n8n and Power Automate as supported automation platforms.

Spaces and permissions

Spaces segment data by team, and a two-layer permission model separates what agents can access from who can use them. The Business plan includes 5 spaces.

Frames

Frames turn agent output into interactive dashboards, reports and memos that a team can explore and work on together.

Slack and Chrome extension

Agents can be mentioned in any Slack channel or messaged directly, and a Chrome extension opens them in a sidebar on any web page.

GPT Researcher

Planner and parallel researchers

A planner splits a question into sub-questions and runs a researcher on each at the same time before a writer drafts the report.

Many retrievers and your own files

The site lists 20+ retrievers, including Tavily, Google, Bing, Exa, arXiv and PubMed, which can be combined with MCP servers or local documents.

Cited reports

Each claim in the report is tied to a numbered source. Reports can be output as markdown, PDF, docx or plain text.

Deep research mode

Findings seed new questions, so a run explores a tree of subtopics instead of a single pass. Depth is configurable.

MCP server and API

The official gptr-mcp server exposes research tools to Claude, Cursor, n8n and other MCP clients, and a FastAPI server can be run from the repository.

Dust

• Free$0

  A Business-plan seat with 500 credits for its lifetime and no monthly refill, meant for occasional users or people trying Dust.

• Pro$24

  per seat / month, billed yearly; $30 billed monthly

  8,000 credits per seat each month on the Business plan, which covers teams of up to 100, with SSO available on request from 5 seats.

• Max$120

  per seat / month, billed yearly; $150 billed monthly

  40,000 credits per seat each month, for people who regularly run complex automations, deep research or tool-heavy workflows.

• EnterpriseCustom

  Unlimited connectors and MCP servers, workspace-pooled credits, SCIM, audit logs, custom data retention, single-tenant deployment and an SLA.

GPT Researcher

• Open Source$0

  MIT-licensed Python package and MCP server that you host yourself; you pay your own LLM and search API providers.

Dust

• Annual discount

  Save 20% with yearly billing

  Dust's pricing page shows a 20% saving on yearly billing: Pro is $24 per seat a month instead of $30, and Max is $120 instead of $150.

GPT Researcher

No public discounts listed.